[PATCH 1/2] data: fix some spelling mistakes in klog and olog databases

[Jeremy Kerr]

Hi Colin,

> diff --git a/data/olog.json b/data/olog.json
> index dd1d25a..9380ba0 100644
> --- a/data/olog.json
> +++ b/data/olog.json
> @@ -22,7 +22,7 @@
>     "pattern": "creating dt node /ibm,opal/power-mgt failed"
>    },
>    {
> -   "advice": "An error condition occured in sleep/winkle engines timer state machine. Dumping debug information to root-cause. OPAL/skiboot may be stuck on some operation that requires SLW timer state machine (e.g. core powersaving)",
> +   "advice": "An error condition occurred in sleep/winkle engines timer state machine. Dumping debug information to root-cause. OPAL/skiboot may be stuck on some operation that requires SLW timer state machine (e.g. core powersaving)",
>     "compare_mode": "string",
>     "label": "SLWRegisterDump",
>     "log_level": "LOG_LEVEL_LOW",

Good catch - took me a while to spot the change with, even looking
straight at the diff :)

The olog.json file is generated from the skiboot firmware source tree
though (which, of course, has the same spelling error):

https://github.com/open-power/skiboot/blob/master/hw/slw.c#L1274

To avoid losing this change next time we re-generate the olog.json data,
we'll fix this in the skiboot tree too.

However, I'd like to ensure that future changes to the fwts tree don't
get out-of-sync with the sources used to generate the json. I have
couple of ideas for that:

 - we only manually update the skiboot tree, and get fixes to the json
   as a result of re-generating the olog.json and sending diffs to fwts
   (so instead of patching fwts, you can tell us we have a typo in
   skiboot, and we'll handle the rest); or

 - we ensure that local changes to olog.json are reviewed by someone
   (either Deb or myself) that can make the corresponding change in
   skiboot

or let me know if you have any other proposal that would work better.

Would it help to add a comment to the top of olog.json, to indicate that
it's generated from other code? Assuming the parser handles comments.

Cheers,


Jeremy